南京工业大学学报（自然科学版）2025，Vol.47Issue(6)：699-706,8.DOI:10.3969/j.issn.1671-7627.2025.06.010
基于BP神经网络和蒙特卡罗法的空间Y形钢箱肋拱桥构件可靠度分析
Reliability analysis of spatial Y-shaped steel box-rib arch bridges based on BP neural network and Monte Carlo method
摘要
Abstract
In the reliability analysis of spatial Y-shaped steel box-rib arch bridges,the structural performance function is highly nonlinear and implicitly expressed.To address this challenge,this study combined backpropagation(BP)neural network with the Monte Carlo(MC)method to evaluate the reliability of bridge components.Test samples were selected by uniform design,and finite element analysis was performed to generate training data for the BP neural network.Based on the trained neural network,Monte Carlo sampling was conducted to estimate the reliability indices of the components.The effectiveness of the proposed method was verified through a numerical example.The results showed that the reliability indices of all major components met the requirements of current design codes,demonstrating good structural robustness.The secondary arch ring exhibited the highest reliability(with reliability indices all greater than 6.0),followed by the main arch ring(with reliability indices all greater than 5.4),while the hangers exhibited the lowest(with reliability indices between 5.3 and 5.7).The reliability of the main arch ring improved after bifurcation,whereas the transition zone between the single and double cable planes(the 10th hanger)represented the weakest link.This study validated the rationality of the bridge design and provided a theoretical basis for identifying weak components in similar nonstandard arch bridges.关键词
空间Y形钢箱肋拱桥/结构功能函数/BP神经网络/蒙特卡罗法/可靠度/均匀设计Key words
spatial Y-shaped steel box-rib arch bridge/structural performance function/BP neural network/Monte-Carlo method/reliability/uniform design分类
交通工程引用本文复制引用
邓志海,邬晓光,平森森,冯清源,陈其达..基于BP神经网络和蒙特卡罗法的空间Y形钢箱肋拱桥构件可靠度分析[J].南京工业大学学报（自然科学版）,2025,47(6):699-706,8.基金项目
陕西省交通运输厅项目(21-62k) （21-62k）
山西省交通运输厅项目(2021-1-1,2021-1-2) （2021-1-1,2021-1-2）
